A massive Microsoft service outage this week left businesses and government agencies scrambling to access critical services — and it landed at the most awkward possible moment for a company cutting thousands of jobs while pouring roughly $190 billion into AI.

The disruption, which hit Thursday, affected Microsoft 365 services including Outlook, Teams and Copilot, according to live updates tracked by the Financial Express. The global Outlook outage forced businesses to improvise, with some reportedly falling back on phone numbers and Bluetooth to transfer recordings.

Fast Company has reached out to Microsoft for comment. The root cause of the disruption remains unknown.

What has Microsoft done to its workforce?

The company's headcount reduction has been unusually aggressive, and unusually well documented. Here's the timeline:

• April: Microsoft extended buyout offers to 9,000 U.S. employees, its first voluntary retirement program for American staff.
• July: It laid off 4,800 workers, mostly in the sales and Xbox divisions, according to Microsoft's own company blog.
• October: Buyout offers went out to 60,000 employees across four countries.
• Last month: The Information reported that leadership in Microsoft's Azure cloud division discussed cutting costs by at least 20%. Microsoft denied the report.

That's the people side. The money side points the opposite way. Earlier this year, Microsoft projected roughly $190 billion in capital expenditures for 2026, driven largely by the enormous computing demands of AI and cloud services.

Did the layoffs cause the outage?

Here's the honest answer: nobody knows yet, and there is no evidence that either the workforce reductions or the AI investments directly contributed to Thursday's disruption.

But the coincidence raises a real question for Microsoft and the wider tech industry. As businesses shrink their workforces and lean harder on AI to boost productivity, are they also risking the reliability of the products they already sell?

The mechanism is straightforward. Workforce reductions eliminate experienced employees who understand how a company's systems actually work — the longtime bugs, the quirks, the tribal knowledge of what to do when something breaks. Even when layoffs don't directly cause technical problems, the loss of institutional knowledge (the undocumented expertise long-tenured staff carry) makes those problems harder to solve.

Doing this while simultaneously reorganizing around a fast-moving new technology compounds the risk. Fewer resources are left over to maintain and troubleshoot existing systems. The challenge is particularly acute for massive tech companies like Microsoft, whose sprawling infrastructure runs decades-old systems right alongside cutting-edge AI services.

Why smaller teams don't guarantee reliability

Microsoft and many other firms are betting that AI lets smaller teams maintain or even increase productivity. That bet may pay off on output. It doesn't automatically translate into reliability.

Keeping complex systems running requires experienced people who can anticipate failures and understand how different components interact — especially mid-crisis, when something has already gone wrong. When 41 leaders were asked what work they'd never hand off to AI, keeping critical operations running ranked among their concerns.

The burden doesn't fall on humans alone. As companies expand AI services and stack new layers of infrastructure on top of old ones, their systems grow more complex and potentially more vulnerable. That can make outages harder to prevent — and harder to contain once they start.

What's at stake for Microsoft's transformation?

While the root cause of this week's disruption remains to be seen, the incident is a blunt reminder of what Microsoft's transformation actually requires. The company has to accomplish more with fewer workers. That's the stated plan.

But it also has to do so without sacrificing the services its customers already depend on — and Thursday showed just how many customers, from businesses to government agencies, sit on the other side of that bargain. Roughly $190 billion in AI spending buys a lot of computing power. It doesn't buy back institutional knowledge, and it doesn't, by itself, keep Outlook online.